[PATCH v1 1/1] PCI / PM: check all fields in pci_set_platform_pm()

[PATCH v1 1/1] PCI / PM: check all fields in pci_set_platform_pm()

[Andy Shevchenko]

When assign new PCI platform PM operations check for all mandatory fields to
prevent NULL pointer dereference.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
---
 drivers/pci/pci.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
index c8b4dbd..badbddc 100644
--- a/drivers/pci/pci.c
+++ b/drivers/pci/pci.c
@@ -530,8 +530,8 @@ static const struct pci_platform_pm_ops *pci_platform_pm;
 
 int pci_set_platform_pm(const struct pci_platform_pm_ops *ops)
 {
-	if (!ops->is_manageable || !ops->set_state || !ops->choose_state
-	    || !ops->sleep_wake)
+	if (!ops->is_manageable || !ops->set_state || !ops->choose_state ||
+	    !ops->sleep_wake || !ops->run_wake || !ops->need_resume)
 		return -EINVAL;
 	pci_platform_pm = ops;
 	return 0;
-- 
2.8.1

Re: [PATCH v1 1/1] PCI / PM: check all fields in pci_set_platform_pm()

[Rafael J. Wysocki]

On Monday, June 06, 2016 05:25:33 PM Andy Shevchenko wrote:
> When assign new PCI platform PM operations check for all mandatory fields to
> prevent NULL pointer dereference.
> 
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

OK in principle, but what's the motivation?

> ---
>  drivers/pci/pci.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> index c8b4dbd..badbddc 100644
> --- a/drivers/pci/pci.c
> +++ b/drivers/pci/pci.c
> @@ -530,8 +530,8 @@ static const struct pci_platform_pm_ops *pci_platform_pm;
>  
>  int pci_set_platform_pm(const struct pci_platform_pm_ops *ops)
>  {
> -	if (!ops->is_manageable || !ops->set_state || !ops->choose_state
> -	    || !ops->sleep_wake)
> +	if (!ops->is_manageable || !ops->set_state || !ops->choose_state ||
> +	    !ops->sleep_wake || !ops->run_wake || !ops->need_resume)
>  		return -EINVAL;
>  	pci_platform_pm = ops;
>  	return 0;
> 

Re: [PATCH v1 1/1] PCI / PM: check all fields in pci_set_platform_pm()

[Andy Shevchenko]

On Wed, 2016-06-08 at 02:17 +0200, Rafael J. Wysocki wrote:
> On Monday, June 06, 2016 05:25:33 PM Andy Shevchenko wrote:
> > When assign new PCI platform PM operations check for all mandatory
> > fields to
> > prevent NULL pointer dereference.
> > 
> > Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> 
> OK in principle, but what's the motivation?

I didn't investigate the guts of the core code, but I'm about to add new
module which will use these facilities. Since the module is slightly
based on existing code for older kernels I noticed those new callbacks.
To prevent potential NULL pointer dereference.


> > ---
> >  drivers/pci/pci.c | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> > index c8b4dbd..badbddc 100644
> > --- a/drivers/pci/pci.c
> > +++ b/drivers/pci/pci.c
> > @@ -530,8 +530,8 @@ static const struct pci_platform_pm_ops
> > *pci_platform_pm;
> >  
> >  int pci_set_platform_pm(const struct pci_platform_pm_ops *ops)
> >  {
> > -	if (!ops->is_manageable || !ops->set_state || !ops-
> > >choose_state
> > -	    || !ops->sleep_wake)
> > +	if (!ops->is_manageable || !ops->set_state || !ops-
> > >choose_state ||
> > +	    !ops->sleep_wake || !ops->run_wake || !ops-
> > >need_resume)
> >  		return -EINVAL;
> >  	pci_platform_pm = ops;
> >  	return 0;
> > 
> 

-- 
Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Intel Finland Oy

Re: [PATCH v1 1/1] PCI / PM: check all fields in pci_set_platform_pm()

[Bjorn Helgaas]

On Mon, Jun 06, 2016 at 05:25:33PM +0300, Andy Shevchenko wrote:
> When assign new PCI platform PM operations check for all mandatory fields to
> prevent NULL pointer dereference.
> 
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> ---
>  drivers/pci/pci.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> index c8b4dbd..badbddc 100644
> --- a/drivers/pci/pci.c
> +++ b/drivers/pci/pci.c
> @@ -530,8 +530,8 @@ static const struct pci_platform_pm_ops *pci_platform_pm;
>  
>  int pci_set_platform_pm(const struct pci_platform_pm_ops *ops)
>  {
> -	if (!ops->is_manageable || !ops->set_state || !ops->choose_state
> -	    || !ops->sleep_wake)
> +	if (!ops->is_manageable || !ops->set_state || !ops->choose_state ||
> +	    !ops->sleep_wake || !ops->run_wake || !ops->need_resume)
>  		return -EINVAL;

This looks OK to me.  platform_pci_run_wake() and
platform_pci_need_resume() assume that if pci_platform_pm is set,
pci_platform_pm->run_wake and pci_platform_pm->need_resume are valid
function pointers.  This is analogous to what we already do for
ops->is_manageable, etc.

Rafael, do you want to take this or should I?

>  	pci_platform_pm = ops;
>  	return 0;
> -- 
> 2.8.1
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-pci" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v1 1/1] PCI / PM: check all fields in pci_set_platform_pm()

[Rafael J. Wysocki]

On Tue, Jul 19, 2016 at 11:37 PM, Bjorn Helgaas <helgaas@kernel.org> wrote:
> On Mon, Jun 06, 2016 at 05:25:33PM +0300, Andy Shevchenko wrote:
>> When assign new PCI platform PM operations check for all mandatory fields to
>> prevent NULL pointer dereference.
>>
>> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
>> ---
>>  drivers/pci/pci.c | 4 ++--
>>  1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
>> index c8b4dbd..badbddc 100644
>> --- a/drivers/pci/pci.c
>> +++ b/drivers/pci/pci.c
>> @@ -530,8 +530,8 @@ static const struct pci_platform_pm_ops *pci_platform_pm;
>>
>>  int pci_set_platform_pm(const struct pci_platform_pm_ops *ops)
>>  {
>> -     if (!ops->is_manageable || !ops->set_state || !ops->choose_state
>> -         || !ops->sleep_wake)
>> +     if (!ops->is_manageable || !ops->set_state || !ops->choose_state ||
>> +         !ops->sleep_wake || !ops->run_wake || !ops->need_resume)
>>               return -EINVAL;
>
> This looks OK to me.  platform_pci_run_wake() and
> platform_pci_need_resume() assume that if pci_platform_pm is set,
> pci_platform_pm->run_wake and pci_platform_pm->need_resume are valid
> function pointers.  This is analogous to what we already do for
> ops->is_manageable, etc.
>
> Rafael, do you want to take this or should I?
>
>>       pci_platform_pm = ops;
>>       return 0;
>> --

I can take it.

Thanks,
Rafael